Red blood cell transfusion is not associated with lymphocyte subpopulation counts in very low birthweight infants

Abstract

BACKGROUND: Very low birthweight (VLBW) infants frequently require red blood cell (RBC) transfusions. Transfusions may modulate immune responses, but their impact on immune development remains incompletely understood. METHODS: 120 VLBW infants participating in the prospective "Immune Regulation of the Newborn Study" (IRON study) were analyzed during the first 42 postnatal days with serial immune phenotyping. Infants were stratified by RBC transfusion exposure. Determinations were assigned to six nominal time points by their actual day of life. Associations were estimated with linear mixed models with a random intercept per infant, with exposure defined both as ever transfused and as the number of red blood cell (RBC) units received before the respective measurement and corrected by the Benjamini-Hochberg procedure. A falsification analysis modelled exposure occurring only after the measurement. RESULTS: 23 of 120 infants received at least one RBC unit within the first 42 days. Transfused infants were more premature, smaller, more anemic and had higher morbidity. RBC transfusion was not significantly associated with alterations in leukocyte subpopulations (e.g. CD8 + T-cell counts (β = -0.161, 95% CI - 0.359 to 0.036, p = 0.11)). No subpopulation reached significance after correction for multiple comparisons. Transfusion occurring only after the measurement gave a coefficient of larger magnitude (β = -0.137) than actual exposure. CONCLUSIONS: RBC transfusion exposure during the first six postnatal weeks was not associated with the counts of any lymphocyte subpopulation examined. The falsification analysis indicates that residual differences reflect the condition of infants selected for transfusion rather than an effect of transfusion.